AFTER IRRADIATION TO THE DEVELOPING BRAIN
Martina Boström
Center for Brain Repair and Rehabilitation Institute of Neuroscience and Physiology
at Sahlgrenska Academy University of Gothenburg
2013
@ Martina Boström Göteborg 2013, Ineko AB ISBN 978-91-628-8772-8 Cover illustration of blood vessels from Mostphotos All illustrations in the introduction of this thesis are made by
Simon Lundholm @ 300Kelvin http://www.300kelvin.se/
This thesis would not exist without
your endless love and support ♥
5
THE DEVELOPING BRAIN
Martina BoströmCenter for Brain Repair and Rehabilitation,
Institute of Neuroscience and Physiology, at Sahlgrenska Academy, University of Gothenburg, Sweden 2013
Abstract
Radiotherapy is commonly used in the treatment of pediatric brain tumors but is unfortunately associated with debilitating negative effects, such as impaired memory and learning. Historically, vascular damage following radiotherapy was considered the primary injury which in turn caused ischemia and necrosis. This hypothesis was supported by studies reporting structural changes to blood vessels, such as thickening of the vessel walls, vessel dilation and enlargement of the endothelial cell nucleus.
Furthermore, quantitative studies observed time- and dose-dependent loss of endothelial cells, vessel length and density after irradiation. Most experimental studies have, however, focused on the mature brain and used high doses of irradiation. Moreover, the adult brain is capable of generating new neurons throughout life in discrete areas of the brain, and these regions are consistently affected by irradiation. Hence, much research has focused on the neural stem and progenitor cells, since a loss of these cell types has been coupled to cognitive decline after irradiation. Other important cell types have therefore been neglected and need to be examined in order to see the full picture.
In this thesis we investigated the effects after a single moderate dose of cranial irradiation (8 - 10 Gy) to the juvenile brain and focused on the vasculature in different areas. Analysis of vascular structure and complexity up to 1 year after irradiation indicated that the vasculature adjusted to the needs of the surrounding tissue. This was observed in both the hippocampus (gray matter) and the corpus callosum (white matter). We did not observe any apparent endothelial cell death, nor any upregulation of genes involved in endothelial cell death acutely after irradiation.
The reduction of neural progenitor cells in the hippocampus was however irreversible and we demonstrated that irradiation in fact accelerated the natural decline in neurogenesis with age. We also investigated the neurovascular niche and found a disruption early after irradiation that however seemed to normalize with time. This hence demonstrated dissociation between the morphological patency of the neurovascular niche and hippocampal neurogenesis.
Using flow cytometry we isolated endothelial cells and investigated gene expression after irradiation. We then surprisingly observed that endothelial cells upregulated proinflammatory genes acutely after irradiation. This has previously not been observed in endothelial cells after in vivo irradiation, but indicates that although endothelial cells seem to be less sensitive to radiation, they are involved in the inflammatory response after irradiation.
Keywords: Radiotherapy, hippocampus, neurovascular niche, neurogenesis, microvessels, endothelial cells
ISBN 978-91-628-8772-8
6
7
Strålbehandling är en vida använd metod för att behandla tumörer i hjärnan och många patienter med primära eller sekundära tumörer får strålning mot hela eller delar av hjärnan. Förbättrade behandlingsprotokoll har lett till en ökad överlevnad hos barn med hjärntumörer, men även resulterat i allt fler biverkningar som drastiskt sänker deras livskvalitet. Mycket talar för att även låga doser av strålning kan leda till svårigheter med minne och inlärning, en effekt som visat sig vara mer uttalad ju yngre patienterna är vid behandlingen och som förvärras med tiden.
Vi använder oss av en modell där unga möss och råttor strålas för att efterlikna de skador och biverkningar som uppkommer hos barn efter strålbehandling. De flesta celler i hjärnan delar sig väldigt sällan och nervcellerna delar sig inte alls. Därför tål hjärnan, som organ betraktat, strålning bättre än många andra organ. Tills nyligen har man trott att hjärnan påverkas av strålning främst genom att små blodkärl skadas och att störd blodtillförsel skulle skada omgivande vävnad. Då strålning framförallt påverkar celler som delar sig, ser vi i denna modell tydliga skador i hjärnans två områden med neurala stamceller. I denna avhandling har vi fokuserat på ett av dessa områden som kallas hippocampus och som anses viktigt för minne och inlärning. Det tros även finnas ett samband mellan minskad nybildning av nervceller i hippocampus och de problem med minne och inlärning som dessa barn uppvisar efter strålbehandling. Förhoppningen är att hitta strategier som skyddar dessa celler och därmed bidra till ökad livskvalitet för barn som överlever sin cancer. Stamceller i dessa områden är inte jämnt utspridda utan är koncentrerade runt blodkärl, sannolikt för att ha konstant tillgång till näring och de signalämnen som transporteras med blodet. Direkt fysisk kontakt mellan stamceller och endotelceller verkar vara
8
har föreslagit att stamcellernas närhet till blodkärl störs efter strålning mot den mogna hjärnan. En central fråga i denna avhandling har därför varit att undersöka om detta sker även i den unga, ännu växande hjärnan.
Studierna i denna avhandling visar att en relativt måttlig stråldos mot den unga hjärnan leder till en bestående minskning av neurala progenitorceller i hippocampus, och att den normala åldersrelaterade minskningen i nybildning av nervceller skedde snabbare i strålade djur. Förutom detta studerade vi även potentiella effekter på blodkärl, och fann att kärldensiteten anpassade sig till den omgivande vävnaden. Detta står i stark motsats mot de minskningar i kärldensitet som tidigare observerats efter strålning, dock efter mycket högre doser än i denna avhandling. Dessutom visar våra resultat att endotelcellerna inte dör direkt efter strålning. Om strålning resulterar i andra effekter än celldöd återstår dock att undersöka. I motsats till tidigare forskning såg vi enbart en akut störning av närheten mellan kärl och celler i hippocampus och att störningen återhämtade sig med tid efter strålning. Detta tyder på att den omogna och mogna hjärnan reagerar olika på strålning, vilket vi redan har sett i andra avseenden. Ett oväntat fynd var att endotelceller verkar vara involverade i det akuta inflammatoriska svaret som sker efter strålning och som har visat sig ha negativa effekter på nybildning av stamceller.
Sammantaget tyder våra resultat på att blodkärlens endotelceller är mindre känsliga för strålning än omgivande celler, åtminstone vid dessa måttliga stråldoser, men att de kan spela en viktig roll i regleringen av den omgivande miljön efter strålning.
9
This thesis is based on the following papers/manuscripts, which will be referred to in the text by their Roman numerals:
I. Irradiation to the young mouse brain caused long-term, progressive depletion of neurogenesis but did not disrupt the neurovascular niche
Boström M, Kalm M, Karlsson N, Hellström Erkenstam N and Blomgren K.
J Cereb Blood Flow Metab 33(6): 935-943 (2013)
II. The hippocampal neurovascular niche during normal
development and after irradiation to the juvenile mouse brain Boström M, Hellström Erkenstam N, Kaluza D, Jakobsson L, KalmM and Blomgren K.
Submitted
III. Irradiation to the young mouse brain impaired white matter growth more in females than in males
Roughton K, Boström M, Kalm M and Blomgren K.
Cell Death and Disease, In press (2013)
IV. Gene expression in endothelial cells isolated by flow cytometry after irradiation to the young mouse brain
Boström M, Kalm M, Hellström Erkenstam N and Blomgren K.
Manuscript
10 + Positive
BBB Blood brain barrier
BLBP Brain lipid-binding protein BrdU Bromodeoxyuridine CB Cerebellum
CC Corpus callosum
CD13 Cluster of Differentiation 13 CD31 Cluster of Differentiation 31 CNS Central nervous system DCX Doublecortin
DG Dentate gyrus
DNA Deoxyribonucleic acid
EC Endothelial cell
ELISA Enzyme-linked immunosorbent assay EPC Endothelial progenitor cell
GCL Granule cell layer Gy Gray
IHC Immunohistochemistry MBP Myelin basic protein
ML Molecular layer
NPC Neural progenitor cell NSC Neural stem cell
NSPC Neural stem/progenitor cell
OB Olfactory bulb
Olig2 Oligodendrocyte lineage transcription factor 2
P Postnatal day
RMS Rostral migratory stream
RNA Ribonucleic acid
SGZ Subgranular zone
SVZ Subventricular zone
11
Abstract ... 5
Populärvetenskaplig sammanfattning ... 7
List of original papers ... 9
List of abbreviations ... 10
Introduction ... 13
Vascular and nervous system ... 15
Vasculature ... 16
Vasculogenesis and angiogenesis ... 16
Adult neurogenesis ... 18
Concept of neurogenesis ... 18
Adult neurogenic areas ... 19
Apoptosis ... 22
Prolonged postnatal neurogenesis in the cerebellum ... 22
Neurovascular niche ... 23
Blood brain barrier ... 24
Endothelial cells ... 26
Pericytes ... 27
Microglia ... 28
Astrocytes ... 29
White matter ... 30
Corpus callosum ... 30
Oligodendrocytes ... 31
Childhood cancers ... 32
Radiotherapy ... 33
Radiobiological effects ... 34
General aim ... 37
Specific aims ... 37
Methodological considerations ... 39
Animals ... 39
Ethical permission ... 39
Irradiation model ... 40
General irradiation procedure ... 40
Clinical perspective ... 42
Tissue preparation ... 44
Staining procedures ... 45
General procedure of immunohistochemistry ... 45
Possible amplification steps ... 46
Fluorescent molecules ... 47
12
BrdU labeling ... 49
Cell cycle markers ... 50
Stereology ... 51
Landmarks ... 51
Volume measurements ... 52
Cell counting ... 52
Vessel morphology ... 53
Manual vessel analysis ... 53
Automated vessel analysis with Metamorph ... 54
Confocal microscopy ... 55
Neurovascular niche ... 56
Pericyte coverage ... 56
Flow Cytometry... 57
Single cell suspension ... 57
Basic procedure of flow cytometry ... 58
Cell cycle analysis ... 59
Gene expression analysis ... 59
RNA purification ... 59
cDNA synthesis ... 60
Reverse transcription quantitative PCR (RT-qPCR) ... 61
Detecting genomic DNA ... 61
ELISA ... 62
Statistics and significance ... 63
Results and discussion ... 65
Permanent changes in neurogenesis ... 65
Structural changes in the vasculature ... 66
Changes in the neurovascular niche ... 67
Endothelial apoptosis ... 69
Gene expression in endothelial cells after irradiation .... 72
White matter damage ... 74
Vascular versus Neuronal hypothesis ... 76
Concluding remarks ... 79
Specific conclusions to given aims ... 79
Clinical perspective and future directions ... 81
Acknowledgements ... 83
References ... 87
13
Introduction
The large number of postmitotic neurons and other cell types with limited proliferation capacity makes the brain relatively radioresistant compared to other tissues with more rapid cell turnover. This view encouraged the use of high radiation doses within the central nervous system (CNS) during the previous century (Russell et al., 1949). However, shortly after the discovery of the X-rays (end of the 19th century) it was noticed that high radiation doses resulted in necrosis and gliosis of the brain some time after radiation. Major blood vessel abnormalities were a consistent finding within the damaged tissue and it was therefore hypothesized that the damage to normal tissue after irradiation was related to insufficient vascular supply (Hopewell et al., 1993). The validity of this hypothesis has been widely challenged but for long there was ample amount of evidence that the vasculature was the primary target that subsequently caused ischemia and necrosis (McDonald and Hayes, 1967).
The survival rates of children with childhood cancers have increased significantly during the last decades (~6) (Steliarova-Foucher et al., 2004, Gustafsson et al., 2013). As a consequence, a population of long term survivors which previously did not exist is now emerging. Individuals treated with cranial radiotherapy gradually develop negative so called late effects, such as hormonal imbalance, perturbed growth as well as impaired learning and memory (Lannering et al., 1990a). During the last decades (~3) treatment protocols have therefore been adjusted to reduce both radiation volumes and doses in the hope of reducing the negative side effects of radiotherapy. In line with this, many animal models have been redesigned, using lower radiation doses in order to more closely mimic the clinical settings. In the current thesis we have therefore exclusively used moderate
14
doses of radiation. This explains, at least partly, why we were unable to see the kind of vascular damage reported in previous studies following irradiation. However, it is remarkable that these relatively moderate doses nearly ablated hippocampal neurogenesis and that the natural, age-related decline in neurogenesis progressed more rapidly after irradiation. The link between cognitive decline and reduced hippocampal neurogenesis demonstrates the importance of finding ways to prevent or ameliorate this chronic ablation of neurogenesis after radiotherapy.
15
Vascular and nervous system
On the macroscopic level, anatomical similarities between the vascular and the nervous system were observed in the 16th century by the Belgian anatomist Andreas Vesalius. Furthermore, when methods such as histology evolved, a mutual relationship between the two systems was found also on the microscopic level. In fact, the vascular and the nervous system rely on reciprocal communication in order for correct development and functional integration to occur (Mukouyama et al., 2002, Larrivee et al., 2009, Tam and Watts, 2010). Today, it is generally accepted that the modern nervous system arose earlier in evolution than the vascular system. The question regarding which of the systems that regulates or coordinates the other is however under debate. In the developing mouse embryo, outgrowth of the vasculature has been observed to precede neural axon outgrowth (Carmeliet and Tessier- Lavigne, 2005, Tam and Watts, 2010). However, at the same time is has been suggested that peripheral nerves in the skin establish branching patterns of blood vessels as well as arterial differentiation (Mukouyama et al., 2002).
A model has emerged where endothelial tip cells in vessels sense and navigate to their target in a similar fashion that axonal growth cones sense the surroundings and find their way (De Smet et al., 2009). In addition, multiple molecules and signal pathways are used by both the neural and vascular system. Several of these molecules were initially discovered as important for axonal pathfinding (e.g. ephrins, netrins, slit and sematophorins) but have also been identified as involved in vascular remodeling and vessel guidance (Autiero et al., 2005, Carmeliet and Tessier-Lavigne, 2005, Eichmann et al., 2005). Likewise, some well-known factors involved in angiogenesis (e.g.
vascular endothelial growth factor; VEGF and angiopoietin 1; Ang-1) are also regulators during development and normal function of the nervous system (Lee et al., 2009). VEGF is in fact a potent stimulator of neurogenesis
16
both in vivo and in vitro (Jin et al., 2002). Hence, there is a close connection between the vascular and the nervous system.
Vasculature
During the evolution of multicellular organisms, the delivery of oxygen to all tissues by simple diffusion became insufficient. The need for a circulatory system increased as multicellular organisms progressively evolved (Fisher and Burggren, 2007). In aerobic animals, the vascular system exhibits crucial functions in all tissues of the body through the delivery of oxygen and nutrients, as well as the removal of waste products. In fact, since the vasculature is essential for the development of all other organs, it is the first system to develop during embryogenesis (Hirschi et al., 2002). The human brain has an extensive vascular network consisting of approximately 400 miles of blood vessels. Through these, the brain receives and uses around 20% of the energy consumed within the body. This is remarkable in the sense that the brain only constitutes 2% of the total body mass. A continuous blood supply to the brain is crucial given that the brain has no local energy reserve and damage to neurons can therefore occur within minutes if cerebral blood flow stops or decreases. Vessel diameter and cerebral blood flow are dependent on the local demand for oxygen and nutrients and is consequently not fixed for a certain area or region. Increased neuronal activity is therefore intimately coupled to increased cerebral blood flow (Lok et al., 2007, Zlokovic, 2008, Tam and Watts, 2010).
Vasculogenesis and angiogenesis
The formation of blood vessels involves two different processes;
vasculogenesis and angiogenesis. In the embryo, new vessels are formed from endothelial progenitor cells (EPCs, angioblasts), which together create a vascular plexus consisting of a meshwork of endothelial cells (ECs). This
17
process of de novo production of blood vessels is known as vasculogenesis and is established before blood flow begins. The vascular plexus then serves as template for further growth and modifications into a mature vascular system. As the embryo grows, expansion of the vascular system is instead dependent on angiogenesis. The term angiogenesis is generally used to describe the growth of vessels, but in its strictest sense only refers to vessel sprouting and elongation from preexisting ECs. Angiogenesis does hence not include de novo generation of vessels (Hirschi et al., 2002, Swift and Weinstein, 2009, Tam and Watts, 2010, Potente et al., 2011).
Vasculogenesis was originally believed to occur exclusively during embryogenesis and instead it was believed that all postnatal formation of new blood vessels occurred through angiogenesis. However, during recent years it has been suggested that EPCs from the bone marrow can participate in postnatal neovasculogenesis (Khakoo and Finkel, 2005, Ribatti, 2007). EPCs were first characterized by Asahara et al. when they isolated CD34+ cells from human peripheral blood and showed that these cells were capable of differentiation into ECs in vitro (Asahara et al., 1997). The EPCs have been shown to play a critical role in both vascular homeostasis and postnatal vasculogenesis (Khakoo and Finkel, 2005). Furthermore, the EPCs can incorporate into sites of angiogenesis in models of ischemia, hence proposing that enhancing EPC-dependent vasculogenesis could be used for therapeutic purposes in pathological situations (Asahara et al., 1997). However, these cells have also been proposed to play an essential role in the abnormal vascular growth occurring in tumors and it has been shown that ablation of EPCs can reduce tumor growth (Mellick et al., 2010). In addition, compared to healthy controls, increased numbers of circulating EPCs have been found in lung cancer patients (Nowak et al., 2010). The potential therapeutic usefulness of EPCs therefore demonstrates opposite purposes in different
18
pathological situations: (i) In ischemic tissue EPCs could be used to increase vessel growth and (ii) in tissues with abnormal vessel growth (such as tumors), EPCs could be targeted in order to inhibit further vessel expansion.
Adult neurogenesis
Concept of neurogenesisThe term stem cell means that a cell is capable of self-renewal and to give rise to a clonal progeny capable of differentiation. To be named neural stem cell (NSC), the cell needs to be capable of producing the three major cell types of the brain: neurons, oligodendrocytes and astrocytes (Figure 1). NSCs can give rise to neural progenitor cells (NPCs) that are restricted to the neuronal lineage (Zitnik and Martin, 2002).
Figure 1. Schematic illustration of the neuronal stem cell lineage in the brain where neural stem cells divide to give rise to new neurons, oligodendrocytes and astrocytes (illustration made by Simon Lundholm @ 300Kelvin).
Neural stem cell
Glial
progenitor cell
Neuron
Oligodendrocyte
Astrocyte Neural
progenitor cell
19
Neurogenesis is the mechanism by which new neurons are generated from neural stem/progenitor cells (NSPCs). This is most pronounced during embryonic/prenatal development when the newly born neurons populate different areas in the continuously growing brain. The existence of neurogenesis in the adult brain has however historically been extensively questioned.
In the beginning of the 20th century, most neuroscientist including Ramon y Cajal, believed that the adult brain had a definite number of cells without regenerative capacity (Stahnisch and Nitsch, 2002). However, this general consensus was later challenged by Joseph Altman when he discovered that the adult rat brain actually did contain newborn neurons (Altman and Das, 1965). The existence of adult neurogenesis has been observed in many mammals but also in other non-mammalian species, e.g. in insects, reptiles and birds (Goldman and Nottebohm, 1983, Font et al., 1991, Cayre et al., 1996). This indicates that adult neurogenesis could represent an evolutionary ancient phenomenon. Compared to neurogenesis in the embryo, neurogenesis in the adult brain is however considerably limited.
Adult neurogenic areas
Today, neurogenesis is accepted to exist in two areas of the adult brain: (i) adult neurogenesis occurs is the subventricular zone (SVZ) located in the walls of the lateral ventricles. Progenitor cells born in the SVZ migrate long distances via the rostral migratory stream (RMS) to the olfactory bulb (OB) where they differentiate into new interneurons (Alvarez-Buylla et al., 2008).
In the current thesis we have however not studied the SVZ but instead focused on (ii) the subgranular zone (SGZ) located at the border between the hilus and the granule cell layer (GCL) of the hippocampal dentate gyrus (DG). Neurogenesis in the SGZ gives rise to new granule cells which make
20
dendritic connections within the molecular layer (ML), followed by axonal connections with the hippocampal CA3 region (Kempermann et al., 2004a, Frielingsdorf and Kuhn, 2007, Grote and Hannan, 2007, Ming and Song, 2011). As much as 80-85% of the granule cells are generated after birth in the rodent DG, (Bayer, 1980). Kempermann et al. have shown that the majority of new granule cells born in the adult rodent brain stay within the inner third of the GCL (Kempermann et al., 2003). This was further supported by the finding that stem cells and their progeny only give rise to about 1% of the granule cells in the outer adult GCL (Lagace et al., 2007). Furthermore, it has been shown that hippocampal neural stem cells have vascular endfeet, a feature normally attributed to astrocytes. It was therefore proposed that these radial glia-like stem cells arise from astrocytes (Filippov et al., 2003), commonly referred to as type-1 cells (Figure 2) (Kempermann et al., 2004a).
This model has assumed that only one type of radial glia-like stem cell exists.
However, it was recently shown that the adult DG has a population of antigenically heterogeneous radial glia-like stem cells (DeCarolis et al., 2013). In 1998, Peter Eriksson discovered that neurogenesis in the SGZ of the DG also occurs in the adult human brain and that neurogenesis is a life- long process (Eriksson et al., 1998). The adult human hippocampus is estimated to generate 700 new neurons daily in each hemisphere.
Furthermore, it was recently shown that the turnover rate of hippocampal neurons does not differ between females and males (Spalding et al., 2013).
21
Figure 2. Development of neural stem cell to mature neuron in the hippocampal dentate gyrus.
A model has proposed that new granule cells in the hippocampus are produced by six developmental milestones (starting with the type 1 cell and ending with a mature neuron) (Kempermann et al., 2004a). The cells express various markers during differentiation and maturation (some of them are outlined in the figure). The hippocampal circuit is usually described as trisynaptic. First, the mature granule cells in the GCL receive input from the entorhinal cortex via the perforant path. The granule cells, in turn, projects axons to CA3 neurons, where the signal is relayed through the Schaffer collateral fibers to the CA1 region.
Finally the output of the hippocampus is relayed back to the entorhinal cortex (illustration made by Simon Lundholm @ 300Kelvin).
22 Apoptosis
Apoptosis or programmed cell death during mammalian development of the brain is an important process to eliminate superfluous or incorrectly functioning neurons (Kuhn et al., 2005). Every day as many as 9 000 new cells are produced through neurogenesis in the SGZ of the DG in a young adult male rat. But if all of these newly generated cells would survive, the number of cells in the GCL of the DG would be twice as great in only a few months. To prevent this, cells must be eliminated. Approximately half of the newly generated cells die during the first weeks of their life, independent of how old the animal is. The majority of dying cells in the GCL are immature neurons undergoing apoptosis when they shift from transiently amplifying progenitor cells to neuroblasts. Consequently, neurogenesis in the DG is probably not a process of replacement but rather refinement of the amount of new cells (Frielingsdorf and Kuhn, 2007, Sierra et al., 2010).
Prolonged postnatal neurogenesis in the cerebellum
Besides the two neurogenic regions, generation of new neurons is restricted to fetal development in most areas of the brain. However, similar to the DG, most interneurons in the GCL of the rodent cerebellum are formed during the first two postnatal weeks (Miale and Sidman, 1961). Furthermore, proliferation in the human cerebellum continues until the 11th postnatal month (Abraham et al., 2001). However, although significant postnatal neurogenesis occurs in the cerebellum it is not considered an adult neurogenic region. The cerebellum was first identified as important for motor functions but has also been linked to higher motor and cognitive functions such as language production and motor planning (Rapoport et al., 2000, Konczak and Timmann, 2007).
23
Neurovascular niche
The concept of a specific vascular niche within the neurogenic areas was first described by Palmer et al. when they showed that newly generated neurons in the hippocampus were associated with dividing ECs (Palmer et al., 2000).
This suggested that NPCs and ECs either respond to the same mitogenic signals or that the cell division of one cell type generates mitogens that in turn triggers the mitotic expansion of the other (Goldman and Chen, 2011).
Hence, further studies were needed in order to verify the functional implication of such a specialized vascular niche. Accordingly, in vitro experiments showed that ECs release soluble factors that are involved in maintaining the multipotency of stem cells, as well as stimulating self- renewal and increasing neuronal production (Shen et al., 2004). When angiogenesis is pharmacologically inhibited by endostatin, decreased angiogenesis is accompanied by reduced neurogenesis (Nih et al., 2012).
Furthermore, stem cells are in close contact with the basal membrane of ECs by extending vascular end feet (Filippov et al., 2003).
When the vascular systems of two mice of different ages were surgically joined together by parabiosis, neurogenesis in the young mouse decreased while neurogenesis in the old mouse increased. A link between age-related reduction in neurogenesis and intrinsic factors in the blood was hence proposed (Villeda et al., 2011). Moreover, the importance of the microenvironment was demonstrated when cerebellar precursors were ectopically transplanted into the DG of neonatal rats, where they integrated into the hippocampal GCL. The transplanted neurons were morphologically similar to hippocampal neurons and expressed the same cell-type specific proteins (Vicario-Abejon et al., 1995). In addition, when striatal progenitor cells were transplanted into cortical environment they adopted features specific for cortical neurons (Fishell, 1995). The microenvironment has
24
consequently been proposed to determine the regional fate of neural precursors.
Blood brain barrier
The blood brain barrier (BBB) is a physical and metabolic barrier that separates the CNS from the systemic circulation. Physiologically, the BBB is mainly composed of three different cellular elements: ECs, astrocytic endfeet and pericytes (Figure 3). However, it is believed that other cell types such as microglia are also involved in the most important and crucial feature of the BBB which is to maintain homeostasis and limit the penetration of pathogens and toxins into the brain. This is maintained by restricted and regulated exchange of different molecules between the blood stream and the brain (Correale and Villa, 2009, Goldberg and Hirschi, 2009, Tam and Watts, 2010). The BBB is established during fetal development and is well functional by birth (Abbott et al., 2010). Vulnerability of the BBB has been proposed to be age-dependent with increased susceptibility to hypoxia ischemia and inflammation in younger animals (Semple et al., 2013).
25
Figure 3. Illustration of the different cell types involved in maintaining the blood BBB together with other important cell types in this thesis. The BBB physically consists of microvessels (red) with inner lining of endothelial cells (purple), surrounding pericytes (green) and astrocytic endfeets (yellow). Microglia (purple) has also been proposed to play an important role in the BBB by the defense against pathogens. Finally, oligodendrocytes (blue) generate myelin sheaths that wrap around neuronal axons (grey). The astrocyte has contact with both vessels and neurons, thereby acting as a key mediator between these two cell types (illustration made by Simon Lundholm @ 300Kelvin).
26
Endothelial cells
The ECs constitute the inner lining of blood vessels and are therefore considered the key component of the BBB. Vascular endothelium is so called semipermeable, which means that the transport into and out of the blood is regulated. However, ECs display an enormous heterogeneity in both morphology and permeability between different organs and a number of physiological properties make the endothelium in the CNS distinct from the vasculature found in the periphery. CNS-derived ECs are connected by tight junctions, lack fenestrations and have low numbers of pinocytotic vesicles (Rubin and Staddon, 1999, Goldberg and Hirschi, 2009, Segura et al., 2009), thereby ensuring a highly regulated exchange between the brain and the blood circulation. Interestingly, transplantation experiments have revealed that BBB properties are not intrinsic to ECs but rather a consequence of signals from the environment within the brain (Stewart and Wiley, 1981).
However, DNA microarrays have revealed different transcriptional profiles of ECs from different tissues and from large vessels and microvessels, thereby suggesting that maintaining a specific endothelial phenotype is not exclusively dependent on environmental signals (Chi et al., 2003).
In humans, ECs have an average lifespan of around 1 year and proliferation of ECs in the mature vascular system is generally considered stable with infrequent cell turnover. However, ECs are not an inert cell type but rather highly metabolic and involved in many crucial processes such as: leukocyte infiltration, permeability, regulating the proliferation and survival of surrounding cells, maintaining homeostatic balance and regulating vasomotor tone (Hirschi et al., 2002, Aird, 2007, Goldberg and Hirschi, 2009).
Originally, the ECs were considered the ultimate compartment of the BBB but nowadays other cell types have also been identified as key components.
27
Pericytes
All vessels have an inner lining of ECs surrounded by perivascular mural cells. Microvessels are associated with solitary pericytes while larger vessels instead are associated with several layers of vascular smooth muscle cells.
During development, ECs of immature vessels secrete growth factors to attract pericytes. The pericytes extend processes that wrap around the wall of microvessels and they are considered important for microvessel stability.
Importantly, ECs and pericytes share the same basement membrane which enables them to communicate directly. Maturation of vessels is associated with attachment of pericytes to the vessel wall (Lindahl et al., 1997, Ramsauer et al., 2002, Bergers and Song, 2005, Lee et al., 2009). The role of pericytes can be compared to that of oligodendrocytes, which are responsible for the myelination of neurons in the CNS; both oligodendrocytes and pericytes support and enhance the function of the cell type that they enclose (Tam and Watts, 2010).
Pericytes are an important compartment of both the neurovascular niche and the BBB. It has been proposed that a mature vascular network consisting of both ECs and smooth muscle cells is crucial for the survival of neuroblasts after cerebral ischemia (Nih et al., 2012). The vasculature in the CNS harbors the highest pericyte coverage compared to other investigated organs. It is interesting to note that increased pericyte density and/or coverage appear to correlate well with barrier properties of the endothelium. In addition, slow endothelial turnover is correlated with high pericyte coverage, and vice versa (Allt and Lawrenson, 2001, Lee et al., 2009, Armulik et al., 2011).
Furthermore, pericytes exhibit contractile properties and can regulate the blood flow at the capillary level by modulating capillary diameter (Peppiatt et al., 2006). Pericyte loss is associated with pathological BBB breakdown, resulting in increased BBB permeability (Armulik et al., 2010, Bell et al.,
28
2010). Interestingly, pericytes have been implicated to act as brain phagocytes at sites with BBB properties (van Deurs, 1976), a role that is normally ascribed to perivascular microglia.
Microglia
Microglia are the main cells of the immune system in the CNS and in the brain represent approximately 10% of the adult cell population. They arise from myeloid cells that during early embryonic development migrate from the yolk sac to the brain. In the brain, microglia progenitors give rise to mature microglia that sustain a population in the adult brain by in situ proliferation. In mice, the majority of microglia are produced during the first two postnatal weeks (Alliot et al., 1999, Ajami et al., 2007). Depending on their activation state, microglia can exhibit both beneficial and detrimental effects on adult neurogenesis (Kohman and Rhodes, 2013). In the so called resting state, microglia have a small cell body with many fine processes and survey the surrounding for signs of damage or infection as well as maintaining homeostasis. However, in response to harmful stimuli microglia undergo several important characteristic changes such as increased proliferation, retraction of processes and swelling of the cell body. In addition, they start releasing several proinflammatory molecules which initiates the inflammatory response (Monje et al., 2002, Kohman and Rhodes, 2013). Furthermore, cocultures of brain ECs and human blood-derived macrophages decrease paracellular permeability, thereby proposing that these cells play an active role in BBB maintenance and physiology (Zenker et al., 2003).
Microglia also play a very important role during development by quickly eliminating excessive newborn cells in the SGZ through phagocytosis.
Importantly, this phagocytosis by microglia does not involve activation of
29
microglia, proposing that phagocytosis is possible without complete microglia activation (Sierra et al., 2010). Cocultures of NPCs and microglia have shown that microglia can enhance postnatal neurogenesis in the SVZ (Walton et al., 2006). Furthermore, a cross-talk between microglia and newly formed neurons has been proposed as beneficial for neurogenesis (Ekdahl et al., 2009).
Astrocytes
Although glial cells are the most abundant cell type in the brain they were originally considered as supportive and passive cell types. Astrocytes account for around 50% of all glial cells in the brain and outnumber neurons by four times in higher mammals. Typically, astrocytes have a stellate shape, thereby the name, with multiple processes extended towards neurons and blood vessels. One single astrocyte can contact several vessels and synapses and it is therefore believed that astrocytes function as key mediators of neurovascular coordination. In fact, it has been proposed that the anatomical position of astrocytes could affect blood flow regulation by delivering signals between neurons and vessels. In addition, astrocytes are involved in the regulation and maintenance of the BBB (Allen and Barres, 2009, Lee et al., 2009, Greene-Schloesser et al., 2012). Astrocytes aid in the guidance of neurons in order for them to make the proper connections during fetal development of the brain (Virgintino et al., 1998). Furthermore, migrating neuroblasts in the RMS are surrounded by a network of astrocytes, referred to as the glial tube and bi-directional signals between the two cell types have been shown to regulate proliferation and migration of the neuroblasts (Cleary et al., 2006).
30
White matter
The CNS is divided into two main components: white matter, which consists mostly of glial cells and myelinated axons, and gray matter, which consists mostly of neuronal cell bodies. White matter was historically viewed as a relatively passive tissue but recent research has shown that an intact white matter is important to maintain cognitive functions such as information processing (Palmer et al., 2012). Interestingly, MRI studies have revealed sexual differences in the density of gray and white matter. Men had a higher density of white matter and women had a higher density of gray matter.
Moreover, men had a higher density of gray matter in the left hemisphere.
For white matter no asymmetry was detected for men, and women showed no asymmetry for neither white nor gray matter (Gur et al., 1999). Furthermore, the general consensus is that the male brain is around 8-10% larger than the female brain. The brain continues to grow after birth and it has been shown that females reach maximum volume earlier than males (10.5 years in females and 14.5 years in males) (Lenroot et al., 2007).
Corpus callosum
The corpus callosum forms an axonal bridge that connects cortical neurons from the two cerebral hemispheres (Figure 4). Functions related to the unification of the hemispheres include e.g. memory retrieval and storage, enhancing language and auditory function. The corpus callosum is the largest white matter commissure in the brain with approximately 200 million fibers, a number that is already set around birth. However, important structural changes such as myelination, pruning and redirection occur during postnatal development (Luders et al., 2010). The corpus callosum continues to grow during childhood and adolescence with the highest growth peak during the first years in life (Lenroot and Giedd, 2006).
31
Figure 4. Sagittal illustration of a mouse brain with specific regions of interest for the current thesis outlined: cerebellum; CB, corpus callosum; CC and dentate gyrus; DG. The illustration also shows the olfactory bulb; OB, rostral migratory stream; RMS and subventricular zone;
SVZ (illustration made by Simon Lundholm @ 300Kelvin).
Oligodendrocytes
Oligodendrocytes generate myelin sheaths that wrap around neuronal axons to increase the neurotransmission within the CNS. The density of oligodendrocytes is remarkably similar between humans and rodents, although of course humans have significantly superior total numbers (Bradl and Lassmann, 2010). However, the density of oligodendrocytes differs between males and females, with the highest densities observed in males. In this aspect it is surprising that the proliferation of new glial cells (predominantly oligodendrocytes and astrocytes) in females is twice as high as the proliferation in males. This could however be explained by the fact that females also have twice the number of apoptotic glial cells. It was therefore proposed that oligodendrocytes in females have a more rapid cell turnover than oligodendrocytes in males (Cerghet et al., 2006, Cerghet et al., 2009).
Historically, myelination in humans was considered to be completed during the first years of life but it is nowadays recognized that myelination continues
SVZ DG
CB CC
OB
RMS
32
into the second and third decades of life (Semple et al., 2013). Interestingly, females have been proposed to have more unmyelinated fibers than males.
However it has been shown that this ratio changes with age. The volume of white matter, myelinated fibers and myelin sheaths is significantly larger in young males compared to young female rodents, followed by the reversed ratio in middle-aged rodents (Cerghet et al., 2009).
Childhood cancers
In Sweden around 50 000 persons per year are diagnosed with cancer. This number is immense compared to the approximately 300 children and adolescents every year diagnosed with malignancies. Although pediatric cancers represent only a fraction of all diagnoses, the increasing number of childhood cancer survivors is creating a public health issue as they transition into adulthood since they today constitute a population which did not exist before. Pediatric cancers are slightly more frequent in boys than girls and are most common in children between 2 and 6 years. (Barncancerrapporten, 2013). After the leukemias, brain tumors are the most common type of pediatric cancers and account for approximately 20-30% of all childhood malignancies. However, it should be noted that malignant neoplasms before the age of 20 are rare (Parkin et al., 1988, Steliarova-Foucher et al., 2004). In both children and adolescents diagnosed with brain tumors, the survival rates have improved significantly during the last decades and the 5-year survival is today more than 80% (Steliarova-Foucher et al., 2004, Gustafsson et al., 2013). However, with the emergence of increased survival rates it became progressively clear that many of the long-term survivors exhibit multiple so called late effects. In fact, a study proposed that as much as 96% of pediatric brain tumor survivors suffer from late effects (Han et al., 2009).
33
Radiotherapy
Cancer in both children and adults is treated with a combination of chemotherapy, radiotherapy and surgery, however the treatment protocols differ significantly. This is due to the fact that children generally tolerate chemotherapy better than adults but are at the same time extremely susceptive to radiotherapy (Barncancerrapporten, 2013), mostly because they are still growing and developing. Cranial radiotherapy is usually not even considered in children younger than 4 years of age. Even very small doses of radiation can therefore cause significant damage in children and may influence cognitive function into adulthood (Hall et al., 2004). Radiotherapy is historically the treatment strategy associated with the most complications, but improved protocols during the last decades have significantly reduced these problems. However, the complications still exhibit a major limitation in the treatment. Since radiotherapy commonly is combined with other treatment modalities it complicates the discrimination of what negative effects are attributed to radiation alone and what effects that are due to chemotherapy, surgery and the disease itself.
Cranial radiotherapy is associated with multiple late effects that last into adulthood such as psychological-emotional dysfunction, intellectual and memory impairments and perturbed growth an puberty (Lannering et al., 1990a, Lannering et al., 1990b, Kuhn and Blomgren, 2011). Young age and female gender are associated with greater negative effects, such as cognitive decline (Ris et al., 2001, Fouladi et al., 2005, Lahteenmaki et al., 2007).
Furthermore, some of the observed effects could be due to white matter damage as younger age at the time of irradiation significantly reduces white matter volume (Reddick et al., 2000, Mulhern et al., 2001, Palmer et al., 2012).
34
Radiobiological effects
During radiotherapy there is a delicate balance between the goal of killing all malignant cells but at the same time sparing normal cells from damage.
Ionizing radiation means that the radiation has enough energy to create ions in its passage through matter. The major effect of ionizing radiation is DNA damage caused either directly or indirectly. In the direct pathway the DNA molecule is directly damaged by irradiation while in the indirect pathway free radicals are created, which in turn cause DNA damage. Both pathways can create single or double strand breaks of the DNA molecule (Gudkov and Komarova, 2003, Lieberman, 2008, Magnander and Elmroth, 2012). Cells are more vulnerable when they are dividing, and since cancer cells divide more rapidly they have higher susceptibility to radiation than normal cells. Normal cells can therefore generally recover from the effects of radiation more easily than cancer cells can. Radiotherapy exploits the small difference in radiosensitivity between tumor cells and normal cells, referred to as the therapeutic index or therapeutic ratio (Dunne-Daly, 1999, Thoms and Bristow, 2010). However, irradiation targets all cells and it is therefore impossible to completely spare the normal tissue. The normal tissue hence becomes the dose-limiting factor during radiotherapy with the goal to damage as few normal and healthy cells as possible.
It is primarily differences in proliferation that account for the different radiosensitivity between different cell types since the biological effects of radiation is usually not visible until the cell tries to divide. Consequently, in high proliferation cells such as tumor cells and normal rapidly dividing cells (e.g. neural progenitor cells); the effects are observed early after irradiation.
However, in non-dividing cells such as postmitotic neurons and rarely diving cells the effects will be observed much later. Following irradiation, a cell can undergo apoptosis, mitotic catastrophe, and permanent or temporary growth
35
arrest while trying to repair the damage. Frequently dividing cells often undergo apoptosis or mitotic catastrophe after irradiation, while less frequently dividing cells to a greater extent either complete DNA repair or enter a stage of growth arrest. Tumor cells undergo mitotic catastrophe to a greater extent than other cell types since they have an unstable genome, lost their regulation of growth, have a dysfunctional repair system and often also lost their capability to undergo apoptosis (Gudkov and Komarova, 2003).
36
37
General aim
The aim of this thesis was to investigate how the neurovascular niche is affected by irradiation to the developing brain. This was done in order to increase the knowledge about how neurogenesis could be restored after irradiation and hopefully also result in beneficial effects on behavior.
Specific aims
I. To investigate structural effects on the vasculature after irradiation in both neurogenic and non-neurogenic areas
II. To investigate the radiation-induced effects on neurogenesis, both short-term and long-term
III. To investigate if and how the neurovascular niche is altered after irradiation
IV. To investigate the molecular response of endothelial cells to irradiation
38
39
Methodological considerations
The following section describes the most important methods used in this thesis. I have tried to highlight pros and cons of each individual method, as well as problems that may occur. For detailed descriptions of the specific experimental procedures, please read the material and methods section for each individual paper/manuscript.
Animals
In the current thesis we only utilized animal experiments (in vivo) as a model of irradiation-induced damaged to the developing brain. Another possibility would have been to use cell culture models (in vitro) which are very useful and sometimes can replace in vivo models partly or completely. However, the main focus on the current thesis was the neurovascular niche that contains many different cell types. Currently used in vitro models of brain vasculature do not recapitulate the in vivo complexity of the neurovascular niche. We therefore chose to use in vivo models only in order to study a complete system where all cell types and physiological responses are included.
Ethical permission
Before conducting animal experiments, all parts of the experimental design need to be approved, previously by the Swedish Animal Welfare Agency (Djurskyddsmyndigheten), currently by the Swedish Board of Agriculture (Jordbruksverket). This procedure ensures that the well-being of the animals is always prioritized. Animals were housed with ad libitum access to food and water and carefully monitored. All experimental designs in this paper were thoroughly reviewed and approved by the ethical committee (applications no. 6/2007, 30/2008, 423/2008, 326/2009 and 361/11).
40
C57BL/6 mice delivered from Charles River Breeding Laboratories in Germany were used in all four papers. In papers I, II and IV, only male mice were used but in paper III we used mice of both sexes to investigate different responses to irradiation between the sexes. If possible, the animals were delivered several days before the irradiation procedure to acclimatize and avoid unnecessary stress that could influence the outcome of the experiment.
Irradiation model
In our model of irradiation we use a linear accelerator with 4 megavoltage nominal photon energy (Varian Clinac 600 CD) located at the Sahlgrenska University Hospital in Gothenburg to irradiate the developing mouse brain.
This accelerator is normally used for radiotherapy of patients (Figure 5). The advantage of using the same irradiation source as that used for radiotherapy of patients is that we come closer to mimicking the clinical settings. The downside is that this adds an extra stress to the animals since they need to be transported by car between the hospital and the animal facility.
General irradiation procedure
In order for the animals to stay completely still during the irradiation procedure, all animals (both control and irradiated) were anesthetized with an intraperitoneal (i.p.) injection of tribromoethanol (Avertin®).
Tribromoethanol is an injectable anesthetic agent that is used during short experimental procedures such as surgery. Repeated use of tribromoethanol is associated with increased morbidity and mortality, which is why tribromoethanol is recommended for single use only. This is one reason why the animals in our studies do not receive fractionated irradiation.
41
The animals were kept on a warm bed (36°C) both before and after irradiation to maintain the body temperature and physiological parameters when separated from the mother. This step is very important, not only to aid survival, but also since hypothermia decreases the radiosensitivity in the tissue and thereby may mask some of the radiation-induced effects. Control animals were anesthetized and kept on a warm bed but did not receive any irradiation.
Figure 5. The irradiation model used in this thesis where the mice were irradiated with a linear accelerator normally used for patients (illustration made by Simon Lundholm @ 300Kelvin).
42
Different radiation qualities (e.g. photons, protons, particles) deposit their energy at different depths during their journey through matter (in our case the brain). In external radiation therapy, a variety of energy sources/radiation qualities can be utilized but the most commonly used for patients are megavoltage photons. For radiation with megavoltage photons, the deposited dose of radiation is low at the skin and the maximum dose is in fact not achieved until several cm into the tissue. This type of irradiation therefore exhibits skin-sparing properties and is very useful in the treatment of many human cancers. However, it is problematic for treatment of superficial lesions in close relation to the skin.
Irradiation with megavoltage photons was used in all papers of this thesis.
This type of irradiation is however problematic for our experimental model of whole brain irradiation (rats and mice) since we want the whole brain to be irradiated with the same dose. Given that maximal dose is deposited several centimeters into the tissue, in our case brains not bigger than a kidney bean, most of the brain would not receive the desired irradiation dose. Therefore, the head is covered with a tissue equivalent material (bolus) to obtain an even radiation dose in the recipient tissue. The bolus material builds up the radiation dose prior to entry of the beams into the brain, and thereby increases the radiation dose deposited at the surface. As a consequence, the maximum dose of radiation is delivered throughout the brain.
Clinical perspective
In all papers/manuscripts in the current thesis, mice were irradiated on postnatal day 14 (P14). Our ultimate aim is to correlate human brain development with the mouse so that the experimental effects we observe could be directly transferred into the clinic. However, there are major
43
differences between human and rodent brain development which significantly complicates this translation (for review see Semple et al., 2013).
When radiation is used for therapeutic purposes, the total dose is divided into different fractions (usually 2 Gy/fraction). This treatment strategy is used to kill cancer cells but at the same time spare normal tissue as much as possible (Kuhn and Blomgren, 2011). It has previously been reported that fractions above 2 Gy should be avoided due to severe tissue damage as a result (Soussain et al., 2009). The optimal scenario would have been to use fractionated doses also in our experimental studies. However, the mice in this thesis were only exposed to single doses of irradiation. This is due to several practical reasons such as the previously mentioned anesthesia issue, stress during transportation and the potential risk of infections when bringing animals back and forth between the animal facility and the hospital. For those reasons, we use the LQ formula to estimate the equivalent dose of 8 and 10 Gy, respectively, if it instead had been delivered in 2 Gy fractions (Fowler, 1989). A single dose of 8 Gy then corresponds to approximately 18 Gy for late effects in normal brain tissue, which is the dose that is delivered for example to the brains and spinal cords of children with relapse leukemia. In addition, a single dose of 10 Gy corresponds to 26 Gy, which is similar to the dose that is delivered to children with medulloblastoma. Compared to the doses that are delivered to the tumor bed of solid brain tumors (up to 60 Gy), the doses we used are therefore considered as relatively moderate doses of irradiation. However, we should always keep in mind that single and fractionated irradiation will have slightly different radiobiological responses.
For example, it has been shown that the molecular response to fractionated irradiation is slower compared to irradiation with a single dose which results in a rapid response (Gaber et al., 2003, Yuan et al., 2006).
44
Tissue preparation
Tissue preparation is one of the most important steps to consider for the optimizing of downstream analysis. The aim is to preserve the 3D structure of the tissue and allow further processing without change. Important objectives to consider when choosing your fixative are therefore preservation of cellular structure, rapid penetration of the fixative into the tissue and avoidance of autofluorescence.
In papers I-III in the current thesis, tissue for histological studies was fixated by transcardial perfusion with physiological saline (0.9%) or PBS followed by perfusion with formaldehyde solution. The rinsing with saline is important since it expels red blood cells from the vasculature, thereby reducing autofluorescence. Furthermore, the perfusion with formaldehyde enables a more rapid fixation of the tissue and thereby increased preservation. To ensure complete fixation, the tissue is also post-fixed in formaldehyde for 24 hours. The advantage of perfusion fixation is apparent when considering that the penetration rate of formaldehyde is around 24 hours to penetrate the center of a 10 mm thick specimen, i.e. approximately 5 mm per day.
However, one major drawback of perfusion with formalin is that fixed tissue does not allow for protein analysis such as ELISA and western blot.
Formaldehyde is the most common fixative for histology studies. In papers I and II, formalin fixation was performed with 4% paraformaldehyde (PFA) solution freshly prepared in the lab. However, due to the hazard risk when preparing and using paraformaldehyde solutions, many laboratories now seek safer alternatives. As a consequence, a solution called Histofix has emerged on the market. Histofix is a buffered 6% formaldehyde solution that eliminates problems associated with toxicity, since it is ready to use without any preparation. In addition, high quality and stability of Histofix enable high
45
reproducibility between different studies. In paper III, we therefore chose to use Histofix instead of PFA.
Staining procedures
General procedure of immunohistochemistry
Immunohistochemistry (IHC) is a method that utilizes the binding of antibodies to detect antigens in a tissue. The binding of an antibody (primary antibody) to its antigen is visualized through different detection systems that can be either direct or indirect and that is visualized by either a fluorescent dye or a chromogenic signal. The direct procedure is quick since the primary antibody is directly conjugated to a label. In comparison, the indirect method utilizes a two-step reaction where the primary antibody is visualized by a labeled secondary antibody (raised against the host species of the primary antibody). The direct method is faster, easily used for multicolor labeling and eliminates the potential nonspecific binding of the secondary antibody arising during indirect labeling. However, the indirect method generally has higher sensitivity than the direct method since the signal is amplified through the use of a secondary antibody that enables the binding of more than one secondary antibody to each primary antibody. Indirect labeling by IHC has been the key method throughout this thesis due to its increased signal amplification.
There are different kinds of primary antibodies. Polyclonal antibodies are usually isolated from serum and are a mixture of antibodies that recognize several epitopes of an antigen. Since the antibody can bind to more than one epitope, polyclonal antibodies can yield increased signal of the target protein.
In addition, due to quick and large quantity production, these antibodies are generally cheap. The other class of antibodies is so called monoclonal antibodies. These antibodies only recognize one epitope and therefore
46
decrease background and potential cross-reactivity. The monoclonal antibodies are produced in smaller quantities and are therefore generally more expensive.
Possible amplification steps
The signal of the staining can be further amplified by utilizing the strong binding between avidin or streptavidin to biotin. In fact, each avidin/strepavidin molecule can bind as many as 4 biotin molecules. When using a biotinylated antibody (primary or secondary), the signal can be amplified by either an avidin-biotin complex (ABC method) or by streptavidin that is conjugated to either an enzyme or a fluorochrome.
For single labeling of different proteins and quantification with light microscope, we have used the ABC kit where a complex of avidin and the enzyme horseradish peroxidase (HRP) is created by incubation prior to addition to the tissue. The avidin-HRP complex is added following incubation with biotinylated secondary antibody. In the final stage, a solution with 3,3´duaminobenzidine (DAB) is added and the HRP enzyme coverts DAB to a colored product (brown). In the current thesis we have also added NiCl2 that makes the color black instead of brown.
In paper III we used streptavidin amplification for visualization of tomato lectin. Neither lectin nor streptavidin are conventional antibodies. Lectins are proteins that bind more or less selectively to carbohydrate-moieties on glycoproteins. In the case of tomato lectin the target is glycophorin and Tamm-Horsfall glycoprotein. Furthermore, streptavidin bind to biotin with high affinity. The primary “antibody” (tomato lectin) was conjugated to biotin and added to the tissue where it bound its “antigen”. This was followed by addition of a secondary “antibody” consisting of streptavidin conjugated to a fluorophore.
47
Fluorescent molecules
Fluorescence is the phenomenon arising through a three-step process that occurs in specific molecules called fluorophores. First, the fluorophore absorbs a photon that results in that an electron changing from the ground state to an excited vibrational state with higher energy. Secondly, the excited molecule collides with other molecules and thereby loses energy until it reaches the lowest excited state. This step is called the excited state life time.
Thirdly, the electron decays back to the ground state and at the same time releases a photon. Since energy is lost when the electron is in its exited state, the emitted photon is always of lower energy (longer wavelength) than the excitation photon. The signal that you visualize in a microscope is the emitted photon.
All fluorophores have characteristic wavelengths for both excitation and emission. The traditional fluorescent dyes (e.g. FITC and PI) have broad fluorescence spectra which makes them difficult to use when doing multicolor labeling. These are also very sensitive to photo bleaching. Photo bleaching is an irreversible process meaning that the fluorescent probe permanently loses its ability to fluoresce. Newer fluorophores such as the Alexa Fluor® dyes instead have narrow spectra and are more photostable. In this thesis, we have therefore almost exclusively used the Alexa fluorophores.
Some antigens are more easily visualized than other. In addition, different fluorophores produce differently strong signals (Alexa 488 is for example generally stronger than Alexa 555 or Alexa 594). In order to optimize the staining and be able to visualize all antigens, it is therefore preferable to combine the weakest antigen with the strongest fluorophore, and vice versa.
With indirect IHC, a relatively small number of secondary antibodies are needed in order for you to decide from analysis to analysis what fluorophore
48
that you want to use in order to detect a certain antigen. However, for the direct IHC you need to label each primary antibody with the desired fluorophore, hence significantly increasing the amount of antibodies needed.
Rapid photo bleaching of the fluorophores can be inhibited by using an antifade mounting media. In the current thesis we used ProLong Gold antifade reagent (with and without DAPI as nuclei stain) which cure within 24 hours, forming a semi-rigid gel. Following this, samples can be stored for months.
Antigen blocking
Non-specific binding of antibodies needs to be prevented by blocking. This is done by adding an excess of blocking agent to saturate any nonspecific binding sites that your antibody could otherwise bind to. The blocking is done prior to the addition of antibodies, and the blocking is present in all antibody steps during the staining procedure. Commonly used blocking agents are bovine serum albumin, gelatin or normal serum from the same animal species that the secondary antibody was raised in. In the current thesis we blocked unspecific binding by the use of donkey serum since all secondary antibodies used were raised in donkey.
Labeling of proliferating cells
In all tissues in our body, new cells are continuously being produced to replace old, damaged or dying cells. This is accomplished by cell division that is intimately regulated during normal homeostasis. New cells are generated from preexisting cells by several defined stages in the cell cycle.
The cell cycle consists of growth phases (G1 and G2), synthesis of DNA (S) and cell division (M). During the cell cycle, the DNA is duplicated and the two copies are then divided into two identical daughter cells. The cell cycle
49
also contains several crucial checkpoints that regulate the continuation or interruption of the cell cycle (Figure 6).
Figure 6. The cell cycle with three important checkpoints outlined. Note that the true duration of the different phases is not correlated to its size within the figure. The M phase is for example very short while longest time is usually spent in the G1 phase (illustration made by Simon Lundholm @ 300Kelvin).
BrdU labeling
Labeling of newborn cells can be accomplished by incorporation of thymidine analogs into the DNA during the S phase of the cell cycle (Kuhn and Peterson, 2008). Thymidine is unique in the sense that it exists in DNA but not in RNA (Kuhn and Cooper-Kuhn, 2007). The incorporated thymidine analogs can later be detected by IHC and thereby allow a “birth-dating” of
Preparation for Mitosis Growth
Growth Preparation
for DNA Synthesis
DNA Replication
G₀ G₁
S M
G₂
G2/MCheckpoint
M
Checkpoint
G1/S Checkpoint TelophaseCytokenesis
- cytoplasmic division
Anaphase
etaphase M ophase Pr
Interphase Cell division
Mitosis - nuclear division
